China introduced temporary rules for generative AI, including a licensing regime

As generative AI use cases see explosive adaptation, China has taken a leading role in determining how rapidly changing technologies should be used, including licensing regimes for service providers.

On Thursday, China’s top cyber regulator inaugurated a provisional set of rules to govern generative AI services, including API providers, serving China-based users.

The question then is whether China’s swift response to rein in generative AI and its strict rules will stifle innovation. Policy makers are acutely aware of those concerns, stressing in the document that the law aims to “balance development and security.”

First and foremost, the law requires generative AI providers to adhere to socialist core values, which ban everything from pornography and terrorism to racism and content that threatens China’s national security.

Algorithms that can influence public opinion, according to the rules, must be registered with the relevant authorities. Generative AI service providers must also obtain administrative licenses according to law, although the document does not say who is required to do so.

When it comes to user protection, the rules stipulate that algorithms may not discriminate based on factors such as ethnicity, gender, age, occupation, or health, and may not be used for anti-competitive behavior. Service providers are encouraged to create anti-addiction systems for underage users, similar to those used in video games.

The service provider is responsible for identifying and stopping the illegal content creation process, and subsequently improving the algorithm and reporting such incidents to the relevant authorities. That means pointing to the creator of the image or chatbot has the potential to create legal issues for individuals.

In addition, regulators have the right to know the specifications of the generative AI model, including training data, size, type, tagging rules, and algorithms.

Finally, AI development in China is a top-down effort. The document calls for the creation of a public data training platform and computing power sharing. Concrete rules have been proposed in Beijing for a state-backed centralized platform that allocates public cloud resources based on customer requirements

Like other key industries, China calls for “self-innovation” in AI algorithms, frameworks, chips, software platforms and other infrastructure, while pushing for “equal and mutually beneficial” international cooperation.

Related Articles

Back to top button